Functionalized analogues of Tr?ger's base: scope and limitations of a general synthetic procedure and facile, predictable method for the separation of enantiomers

Article abstract of DOI:10.1016/j.tet.2008.04.111

A major stumbling block in the applications of enantiomerically pure Tr?ger's base analogues is their poor availability. We have therefore developed a facile method for the enantioseparation of functionalized Tr?ger's base analogues possessing various sub

Full text of DOI:10.1016/j.tet.2008.04.111

Tetrahedron 64 (2008) 6252–6262
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
¨
Functionalized analogues of Troger’s base: scope and limitations of a general
synthetic procedure and facile, predictable method for the separation
of enantiomers
a
a
a
a
ˆ
Delphine Didier , Benoıt Tylleman , Natacha Lambert , Christophe M.L. Vande Velde ,
Frank Blockhuys b, Alain Collas b, Sergey Sergeyev a
,
*
a
´
`
Universite Libre de Bruxelles (ULB), Laboratoire de Chimie des Polymeres, CP 206/01, Boulevard du Triomphe, 1050 Bruxelles, Belgium
b Department of Chemistry, University of Antwerp, Universiteitsplein 1, 2610 Wilrijk, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
A major stumbling block in the applications of enantiomerically pure Tro¨ger’s base analogues is their
poor availability. We have therefore developed a facile method for the enantioseparation of function-
alized Tro¨ger’s base analogues possessing various substitution patterns. The systematic separation of
a library comprising 36 representatives on the commercially available Whelk O1 chiral stationary phase
provided valuable information on structure–enantioselectivity relationships. A mechanistic explanation
of observed relationships allows one to predict whether or not enantioseparation of a given, perhaps yet
unknown derivative of Tro¨ger’s base will be feasible. In addition, we provide a detailed report on the
scope and limitations of the general synthetic protocol employing anilines and paraformaldehyde in
CF3COOH, as well as some considerations concerning the mechanism of formation of Tro¨ger’s base
analogues.
Received 26 March 2008
Received in revised form 24 April 2008
Accepted 28 April 2008
Available online 1 May 2008
Keywords:
Tro¨ger’s base
Enantioseparation
Whelk O1
Anilines
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
base skeleton, and deal with racemates. The advantages of Tro¨ger’s
base chirality thus remain largely unexploited.
Tro¨ger’s
base,
(ꢀ)-2,8-dimethyl-6H,12H-5,11-methanodi-
A major stumbling block in the applications of enantiomerically
pure Tro¨ger’s base analogues is their poor availability. The resolu-
tion of Tro¨ger’s base derivatives with the aid of chiral acids was for
a long time considered as unfeasible due to the acid-promoted
racemization via the formation of an iminium intermediate. There
are a few notable exceptions to this rule,12 and recent publications
reveal considerable controversy in this issue. Thus, Tro¨ger’s base
was successfully resolved with the aid of (ꢁ)-(R,R)-dibenzoyltar-
taric acid in acetone to give (ꢁ)-(R,R)-1a with 91% ee.13 Kostya-
novsky and co-workers,14 as well as Wa¨rnmark and co-workers15
reported a considerable increase in the stability toward racemiza-
tion of Tro¨ger’s base analogues bearing substituents in the ortho-
positions relative to the nitrogen atoms, though the resolution of
benzo[b,f][1,5]diazocine ((ꢀ)-1a, Fig. 1), was first synthesized in
1887 by the condensation of para-methylaniline with formalde-
hyde. Tro¨ger’s base is a chiral diamine with two stereogenic
bridgehead nitrogen atoms. The unique set of structural features
(C2-symmetry and a rigid V-shape geometry with the two aromatic
rings nearly perpendicular to each other) makes derivatives of
Tro¨ger’s base very attractive for applications in supramolecular
chemistry and molecular recognition.1,2 For instance, various
functional groups capable of forming hydrogen bonds were in-
stalled at the extremities of the V-shaped skeleton of Tro¨ger’s base
to create synthetic receptors for the recognition of adenine, pyr-
imidine, biotin derivatives,3 and dicarboxylic acids.4 The rigid
scaffold of Tro¨ger’s base was used in ‘molecular torsion balances’ for
the quantification of weak molecular forces,5 in the templated
synthesis of fullerene derivatives,6–8 and in the synthesis of self-
assembled metallohelicates9,10 or metallomacrocycles.11 However,
many of these applications explore only the geometry of Tro¨ger’s
* Corresponding author. Tel.: þ32 2 6505392; fax: þ32 2 6505410.
E-mail address: sserguee@ulb.ac.be (S. Sergeyev).
Figure 1. Tro¨ger’s base: structural formula (left) and optimized geometry of (S,S)-en-
antiomer (right).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.04.111

D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
6253
such derivatives with the aid of chiral acids was not demonstrated.
Overall, the resolution of Tro¨ger’s base and its analogues via di-
astereoisomeric salts does not appear impossible, but it certainly
requires a careful choice of resolving agents and experimental
conditions on a case-to-case basis, and therefore lacks generality.
Crystallization of racemic conglomerates followed by manual sep-
aration of crystals of enantiopure Tro¨ger’s base analogues16,17 is
fundamentally important, but is of limited practical utility.
As early as in 1944 Prelog and Wieland have separated enan-
tiomers of Tro¨ger’s base by column chromatography on lactose.18
Later on, Tro¨ger’s base became a popular model substance for
various chiral chromatography techniques.19 Currently, both en-
antiomers of Tro¨ger’s base are commercially available. However,
enantioseparations of Tro¨ger’s base analogues are rather scattered
and have never been performed on a systematic basis.16,20,21 In
support of our own studies on asymmetric organocatalysis, we
became interested in enantiopure Tro¨ger’s base analogues bearing
various functional groups in different positions of the meth-
anodibenzo[b,f][1,5]diazocine system. Here, we describe a facile
and predictable method for the enantioseparation of Tro¨ger’s base
analogues by chromatography on the commercially available chiral
stationary phase (CSP) Whelk O1. As the thorough evaluation of
such a method necessitated a relatively large library of racemic
compounds, we also investigated in detail the scope and limitations
of a general procedure for the synthesis of Tro¨ger’s base analogues
from the corresponding anilines.
base.27–30 A few other reported examples, in which this protocol
was successfully applied for the condensations of some other aro-
matic amines,9,31,32 prompted us to choose it for our work and to
investigate systematically its scope and limitations. We therefore
performed condensations of anilines 4–6 bearing various func-
tional groups and having different substitution patterns of the ar-
omatic ring. The outcomes of these reactions are summarized in
Tables 1–3. All reported yields of Tro¨ger’s base analogues 1–3 are
those of analytically pure products after purification by column
chromatography (see Section 4).
For 4-substituted and 2,4-disubstituted anilines with moder-
ately electron-donating alkyl substituents we obtained yields of the
corresponding Tro¨ger’s base analogues (ꢀ)-1a,c–f and (ꢀ)-2a,
which are superior to those for any alternative methods. 4-
Methoxy- and 4-methylsulfanyl anilines 4g,h also gave high yields
of the expected products. A notable contrast was 4-methoxy-2-
methylaniline 5d: an additional methyl group resulted in the dra-
matic drop of yield of the corresponding Tro¨ger’s base analogue
(from 66% for (ꢀ)-1g to 10% for (ꢀ)-2d). This is in agreement with
the earlier reported 14% yield of (ꢀ)-2d from 5d and formaldehyde
in EtOH/H2O/HCl.24 Presumably, when the aromatic ring is too
much activated by electron-donating groups, it becomes prone to
hydroxymethylation via the attack of protonated formaldehyde,
which finally leads to polymeric products in a process that closely
resembles the acid-catalyzed formation of phenol–formaldehyde
Table 1
2. Results and discussion
Condensation of 4-substituted anilines with (CH2O)n in CF3COOH
N
(CH2O)
¨
NH2
2.1. Preparation of Troger’s base analogues from anilines
n
R
CF3COOH
R
R
For our studies on the enantioseparation of Tro¨ger’s base ana-
logues, we needed a relatively large library of racemates. Therefore,
we were interested in a general method, with very clearly estab-
lished scope and limitations, which would permit easy synthesis of
as many as possible Tro¨ger’s base analogues according to a stan-
dard, operationally simple recipe, with minimal, if any, variation of
reaction conditions.
The most used general approach for the synthesis of Tro¨ger’s
base derivatives is the condensation of anilines with formaldehyde
or its synthetic equivalent in the presence of a Brønsted acid.
However, a variety of conditions were applied with very uneven
success. Various sources of reactive formaldehydes were used, as
were different acidic reaction media. Very recent reports on the
Lewis acid catalysis in the synthesis of Tro¨ger’s base,13 or on the
preparation of heteroaromatic Tro¨ger’s base analogues in ionic
liquids22 evidence that there is still an ongoing search for improved
synthetic protocols.
To the best of our knowledge, two procedures for the synthesis
of Tro¨ger’s base analogues were rigorously examined for their
scope and limitations on a relatively large number of molecules.
Wilcox and co-workers found that only anilines with electron-do-
nating substituents afford Tro¨ger’s base analogues in the conden-
sation with CH2O in HCl/EtOH/H2O.23,24 On the contrary, a method
that uses DMSO in AcOH/HCl as an unusual equivalent of formal-
dehyde tolerates a range of functional groups, including strongly
electron-withdrawing substituents. However, yields are often low,
and the use of corrosive gaseous HCl (in some instances, at elevated
temperatures) renders this method rather impractical.25
It was accepted that non-aqueous, highly ionizable reaction
media are in general beneficial for the condensation of anilines
with formaldehyde. Performing condensation of some anilines in
AcOH/HCl instead of EtOH/H2O/HCl considerably improved yields
of a few Tro¨ger’s base analogues.26 Later on, the paraformaldehyde/
CF3COOH method introduced by Wa¨rnmark became popular for the
synthesis of synthetically valuable halogen derivatives of Tro¨ger’s
N
( )-1a–t
4a–t
Tro¨ger’s base
analogue
R
Yield,a
%
Previously reported yields,b
%
(ꢀ)-1a
(ꢀ)-1b
(ꢀ)-1c
(ꢀ)-1d
(ꢀ)-1e
(ꢀ)-1f
(ꢀ)-1g
(ꢀ)-1h
(ꢀ)-1i
(ꢀ)-1j
(ꢀ)-1k
(ꢀ)-1l
Me
H
Et
i-Pr
89
70c
86
86
88
60 (A),24 77 (B),26 55 (D)25
0 (A),24 53 (D)25
t-Bu
n-Hexyl
MeO
MeS
Me2N
CH2CH2OH
CH2OH
F
79
66
69
19.5 (A),33 53 (B),26 60 (D)25
59 (B)26
<10d
54
0 (B)26
46 (A)24
22
0 (A)24
69c
0 (A),26 5–10 (B),26 29–61
(C),27,29 61 (D)25
(ꢀ)-1m
(ꢀ)-1n
(ꢀ)-1o
(ꢀ)-1p
(ꢀ)-1q
(ꢀ)-1r
(ꢀ)-1s
(ꢀ)-1t
Cl
Br
I
57c
55c
56c
74
0 (A,B),26 63–67 (C),27,29 45 (D)25
0 (A,B),26 55–65 (C)27,29
37–64 (C)6,27,29
COOEt
CF3
4-CN-C6H4
CN
234 (E), 37 (D)25
54, 42e
20 (D)25
43f
0
21 (D)25
23 (D)25
NO2
0
a
If not stated otherwise: aniline 4a–t (5 mmol) and then (CH2O)n (10 mmol) were
added to CF3COOH (10 mL) at ꢁ15 ꢂC and then allowed to react at rt for 24 h.
b
Reactions of anilines with the following combinations of reactants and solvents:
A: CH2O/H2O/EtOH/HCl; B: (CH2O)n/AcOH/HCl; C: (CH2O)n/CF3COOH; D: DMSO/
AcOH/HCl; E: (CH2)6N4 in CF3COOH.
c
Compounds (ꢀ)-1b,l–o were synthesized by this method previously under
various conditions.27,29,32 However, the yields of halogen derivatives (ꢀ)-1l–o were
shown to be significantly dependent on the temperature, the scale of reaction, and
the order of the addition of reactants.29 We therefore preferred to reproduce these
syntheses in exactly the same conditions as for the other derivatives.
d
Estimated yield, only impure material was obtained.
Addition of reactants to CF3COOH at ꢁ15 ꢂC followed by heating to reflux
e
for 15 h.
f
Prepared from 2.5 mmol of 4r and 5 mmol of (CH2O)n.

6254
D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
Table 2
resins (see Scheme 1). In accord with this was also the outcome of
the reaction with 4i that bears a very strong electron-donating
Me2N group. In this case, a complex mixture of products was
formed, which did apparently contain some amount of Tro¨ger’s
base analogue (ꢀ)-1i, according to the spectral data of the reaction
mixture after work up.36 Column chromatography afforded a frac-
tion considerably enriched in this material, but all attempts to
isolate pure (ꢀ)-1i failed.
Condensation of 2,4-disubstituted anilines with (CH2O)n in CF3COOH
R1
R1
R2
N
(CH2O)
NH2
n
CF3COOH
R2
R2
N
R1
( )-2a–g
5a–g
An unprotected hydroxyl group is compatible with the reaction
conditions: thus, diol (ꢀ)-1j was prepared from 4-(2-hydroxy-
ethyl)aniline 4j in approximately the same yield as by the con-
densation in EtOH/H2O/HCl.24 Notably, we succeeded also in the
preparation, albeit in modest yield (22%), of diol (ꢀ)-1k, which was
reported to be inaccessible by condensation in aqueous medium,
probably due to the easy formation of a benzyl-type cation that
then attacks another aniline molecule to finally afford polymeric
products in a sequence of electrophilic substitutions similar to that
depicted in Scheme 1.24 This observation is of a certain practical
value, considering the synthetic versatility of the hydroxyl function.
Despite the modest yield, the one-step synthesis of (ꢀ)-1k from
inexpensive 4-aminobenzylalcohol 4k appears competitive with
the known multi-step procedure.6,8
Tro¨ger’s base
analogue
R1
R2
Yield,a
%
Previously reported yields,b
%
(ꢀ)-2a
(ꢀ)-2b
(ꢀ)-2c
(ꢀ)-2d
(ꢀ)-2e
(ꢀ)-2f
(ꢀ)-2g
Me
Br
Br
Me
Me
NO2
Br
Me
Me
CF3
MeO
NO2
Me
Br
87
32–40 (A),24 40 (B),26 57 (D)25
98c
34d
10
14 (A)24
80 (C)9
18, 22,e 81f,g
0e,h
20, 27,i 33g
60j (C)34
a
If not stated otherwise: aniline 5a–g (5 mmol) and then (CH2O)n (10 mmol) were
added to CF3COOH (10 mL) at ꢁ15 ꢂC and then allowed to react at rt for 24 h.
b
Reactions of anilines with the following combinations of reactants and solvents:
A: CH2O/H2O/EtOH/HCl; B: (CH2O)n/AcOH/HCl; C: (CH2O)n/CF3COOH; D: DMSO/
AcOH/HCl.
c
Yield from Ref. 20.
d
Prepared from 2 mmol of 5c and 4 mmol of (CH2O)n.
Addition of reactants to CF3COOH at ꢁ15 ꢂC followed by heating to reflux for
Next to anilines 4l–o,27 also anilines 4p,q bearing stronger
electron-withdrawing groups (CF3, COOEt) gave good yields of the
corresponding Tro¨ger’s base analogues (ꢀ)-1p,q. However, yields of
Tro¨ger’s base analogues with two electron-withdrawing groups per
ring were somewhat lower. Thus, having either CF3 and bromine or
two bromines on the aromatic ring of anilines 5c,g, respectively,
considerably lowered yields of the corresponding Tro¨ger’s base
analogues (ꢀ)-2c,g compared to (ꢀ)-1n. At the same time, 2,5-
dibromoaniline 6a gave 61–67% of the corresponding Tro¨ger’s base
analogue (ꢀ)-3a. Higher yield of (ꢀ)-3a compared to its isomer
(ꢀ)-2g is likely due to the cooperative orienting effect of NH2 and
5-bromo substituents in 6a.
e
24 h.
f
Compound (ꢀ)-2e was synthesized by this method previously. In this paper we
discuss some variations of the reaction conditions. For the sake of unambiguous
comparison, we have also reproduced the original procedure as published.9
g
Addition of reactants to CF3COOH at rt followed by reaction at rt for 24 h.
Formamide derivative 9 (39%) was isolated, see Scheme 3.
Reaction at rt for 21 days.
We are unable to reproduce the recently published 60% yield of (ꢀ)-2g. Although
h
i
j
the molecular structure of (ꢀ)-2g was unambiguously proven by Try and co-workers
with the aid of X-ray diffraction, no other analytical data were published.34 This
raises doubts about either the homogeneity of the material isolated by authors, or
the accuracy of the published yield.
X-ray diffraction analysis of (ꢀ)-3a revealed the angles between
the two planes of the aromatic rings to be 106.4ꢂ and 105.2ꢂ, re-
spectively, for the two crystallographically independent molecules
(Fig. 2; see Supplementary data for detailed discussion of the
crystal structure of (ꢀ)-3a). These values are comparatively large,
but not unique for multiply substituted Tro¨ger’s base analogues.2,34
Also an unprecedented octafluoro analogue of Tro¨ger’s base
(ꢀ)-3b was synthesized from 1,2,3,4-tetrafluoroaniline 6b, al-
though in a modest yield (37%). Considering current interest in
‘oligo-Tro¨ger’s bases’ as concave aromatic receptors (molecular
clips or tweezers),2,15,37 (ꢀ)-3b can be viewed as an initial entry to
a new generation of concave molecules with highly fluorinated
aromatic rings.
Tetrabromides (ꢀ)-2g and (ꢀ)-3a are interesting intermediates
for metal-catalyzed cross-coupling reactions, which were exten-
sively used earlier for the transformation of dibromo and diiodo
analogues of Tro¨ger’s base into variety of complex derivatives.38,39
We attempted to optimize the yield of tetrabromide (ꢀ)-2g. How-
ever, even very long reaction times improved the yield only slightly.
The best, even if still rather modest yield of (ꢀ)-2g (33%) was
Table 3
Condensation of variously substituted anilines with (CH2O)n in CF3COOH
R1
R4
R1
R2
R3
R3
R2
R2
R3
NH2
(CH2O)
N
n
CF3COOH
N
R4
R4
R1
( )-3a–g
6a–g
Tro¨ger’s base analogue
R1
R2
R3
R4
Yield,a
%
(ꢀ)-3a
(ꢀ)-3b
(ꢀ)-3c
(ꢀ)-3d
(ꢀ)-3e
(ꢀ)-3f
(ꢀ)-3g
Br
F
H
F
H
Me
Me
H
H
F
Br
F
H
61, 67b
37c
75
94
71
Me
Me
H
Me
MeO
H
H
H
H
H
H
Me
Me
H
47
H
0
a
If not stated otherwise: aniline 6a–g (5 mmol) and then (CH2O)n (10 mmol) were
added to CF3COOH (10 mL) at ꢁ15 ꢂC and then allowed to react at rt for 24 h.
b
Addition of reactants to CF3COOH at rt followed by reaction at rt for 24 h.
Prepared from 2 mmol of 6b and 4 mmol of (CH2O)n, yield from Ref. 35.
c
Scheme 1. Formation of polymeric products from strongly activated anilines. Positions designated with ‘ ’ are also highly activated toward electrophilic substitution when R and/or
*
R0 are very strong electron-donating groups such as MeO, Me2N.

D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
6255
Table 4
Products from the condensation of 2,4-dibromoaniline with (CH2O)n in CF3COOH
Br
N
Br
Br
Br
N
Br
Br
Br
( )-2g
Br
(CH2O)
NH2
n
NHMe
CF3COOH
Br
Br
H
N
5g
8
N
Br
Br
7
Temperaturea (ꢂC)
Time
(ꢀ)-2g,b
20
33
27
%
7,b
%
8,b
4.5
6
18
%
ꢁ15 ꢂ
rt.
C
24 h
48 h
21 days
51
28
0
ꢁ15 ꢂ
C
a
Temperature at which the reaction was initiated.
Yields of isolated, analytically pure substances.
b
steps still proceed relatively easily. However, when the starting
aniline bears electron-withdrawing groups, they reduce the nu-
cleophilicity of the secondary amine nitrogen in F and slow down
the intramolecular electrophilic substitution in G. An important
side reaction of a relatively unstable F is an oxidation to the dihy-
droquinazoline I that was in some instances isolated earlier.26,40
The following explanation of why analogues of Tro¨ger’s base
bearing moderately electron-withdrawing substituents (e.g., halo-
gens) can easily be prepared in CF3COOH, but not in HCl/H2O, was
suggested by Wa¨rnmark and co-workers:27 in CF3COOH, the con-
centration of the reactive protonated formaldehyde, which is nec-
essary for the transformation of tetrahydroquinazoline F into the
cationic intermediate G is higher than in aqueous HCl, and it sub-
stantially accelerates the rate-limiting step.
We provide somewhat different rationalization based on the
data from Table 4. Isolation of 51% of tetrahydroquinazoline 7
undoubtedly confirms that its further conversion is indeed the rate-
limiting step. However, prolonged reaction times only cause mar-
ginal increase in the yield of Tro¨ger’s base analogue (ꢀ)-2g ((ꢀ)-H,
R¼R0¼Br), while tetrahydroquinazoline 7 (F, R¼R0¼Br) was com-
pletely consumed. Considering that formaldehyde is taken in ex-
cess (2 equiv per 1 equiv of aniline, while 1.5 equiv is required for
the formation of Tro¨ger’s base framework), and that the formation
of tetrahydroquinazoline F only requires 1:1 stoichiometry aniline/
CH2O, there should be sufficient formaldehyde during the entire
course of the reaction. This suggests the participation of a different
Figure 2. ORTEP plot of the asymmetric unit of (ꢀ)-3a with the displacement
ellipsoids drawn at the 50% probability level.
obtained when the reactants were added to CF3COOH in one por-
tion at rt. In an attempt to rationalize the rather low yields of
Tro¨ger’s base analogue (ꢀ)-2g, we decided to carefully isolate side
products and intermediates of this reaction (Table 4). The only
isolable products were tetrahydroquinazoline 7 and N-methyl-2,4-
dibromoaniline 8. Structure of 7 was unambiguously confirmed by
X-ray diffraction analysis (Fig. 3). To the best of our knowledge, no
X-ray structure of tetrahydroquinazolines has so far been reported,
possibly due to the instability of such compounds toward oxidation
(see discussion below).
The accepted mechanism for the formation of Tro¨ger’s base
analogues from anilines and formaldehyde involves the following
key steps: (1) the transformation of aniline A into the iminium
intermediate B that may exist in equilibrium with unstable methy-
lene imine C and diarylaminomethane D; (2) the electrophilic attack
of B at the ortho-position of aniline A to give ortho-aminobenzyl-
amine E; (3) the cyclization of E to give tetrahydroquinazoline F; (4)
the formation of iminium intermediate G from F and finally, the
intramolecular electrophilic substitution in G to give a Tro¨ger’s base
analogue (ꢀ)-H (Scheme 2).
Previously, convincing evidence was obtained that the rate-
limiting steps in the formation of Tro¨ger’s base derivatives are the
conversion of tetrahydroquinazoline derivative F to the reactive
intermediate G and the intramolecular electrophilic substitution in
the latter. When aromatic rings bear activating substituents, these
Figure 3. ORTEP plot of 7 with the displacement ellipsoids drawn at the 50% proba-
bility level.

6256
D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
Scheme 2. Accepted mechanism for the formation of Tro¨ger’s base analogues from anilines and some possible side processes.
electrophile, namely, iminium cation B in the conversion of F to G.
The possibility of this process was clearly demonstrated by Wagner
in his comprehensive study of reactions between aromatic amines
and formaldehyde.41 In the early stages of the reaction, there are
sufficient amounts of both tetrahydroquinazoline F and iminium
cation B, therefore the reactive intermediate G is formed relatively
quickly and then easily cyclizes to give the final Tro¨ger’s base an-
alogue. Later on, the only source of B is the reverse reactions of
intermediates E and F (reversibility of the formation of E and F was
demonstrated earlier).41 As these reverse reactions are apparently
rather slow, the concentration of the reactive cation B becomes low.
Hence, formation of the intermediate G proceeds only sluggishly.
Instead, a considerable amount of B is reduced to N-methylaniline 8
the best result when reactants were added to CF3COOH in one
portion at rt. When reactants were carefully added to CF3COOH
upon cooling to ꢁ15 ꢂC, and the reaction was allowed to proceed
either at rt or at reflux of CF3COOH, the yield of Tro¨ger’s base an-
alogue (ꢀ)-2e was much lower.42 Thus, high concentrations of re-
active intermediates at the earlier stages of reaction appear to be
vitally important in the case of deactivated anilines.43
It is remarkable that the electron-donating methyl group in
aniline 5e counterbalances the presence of the very strong elec-
tron-withdrawing NO2 group, and the formation of Tro¨ger’s base
analogue proceeds efficiently. When strong electron acceptors such
as CN or NO2 are present alone (anilines 4s,t), they deactivate the
aromatic ring to such an extent that no Tro¨ger’s base analogue is
formed. However, 4-methyl-2-nitroaniline (5f) also gave no Tro¨g-
er’s base analogue in any of the discussed modifications of condi-
tions. Instead, the formamide derivative 9 was unexpectedly
isolated when the reaction was conducted upon heating. The
structure of 9 was unambiguously confirmed by X-ray diffraction
analysis (Fig. 4), by NMR, and by ESIMS data. The formation of 9
(J, R¼R0¼Br) in
a hydride-transfer process, well known as
Eschweiler–Clarke methylation, in which CH2O serves as a hydro-
gen donor. When F is completely consumed, N-methylaniline J
becomes the second major product of the reaction. The fact that
after 21 days only 45% of total material can be isolated (vs ca. 76%
after 24 h) clearly indicates also the formation of non-isolable,
probably polymeric products.
It should be noted that N-methyl derivatives of the starting
anilines were observed in virtually all reactions. Thus, in the course
of syntheses of (ꢀ)-1o and (ꢀ)-1q we have isolated 4% of N-methyl-
4-iodoaniline and 4% of N-methyl-4-(trifluoromethyl)aniline, re-
spectively. In other cases, N-methylanilines were not isolated, but
their formation was evidenced by TLC (relatively unpolar spots with
Rf close to that of the corresponding starting aniline) and by 1H
NMR (characteristic signals of MeN at ca. 3.0 ppm).
The fact that the best yield of (ꢀ)-2g from 5g was obtained upon
mixing the reactants at rt is also in good agreement with our
speculations. In this case, the highest concentrations of both B and
F, which are involved in the rate-limiting step, are probably ach-
ieved soon after the start of the reaction. Also in line with this
rationalization, condensation of 2-methyl-4-nitroaniline 5e gave
Figure 4. ORTEP plot of 9 with the displacement ellipsoids drawn at the 50% proba-
bility level.

D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
6257
Scheme 3. Formation of an unexpected product 9 from aniline 5f.
could be explained by the reaction of the initially formed ortho-
aminobenzylamine (Scheme 2, E, R¼Me, R0¼NO2) with formalde-
hyde to give an unstable N-hydroxymethyl derivative 10, which
then undergoes formal dehydrogenation to give the formamide 9.
N-Hydroxymethyl amines such as 10 are probable intermediates in
the formation of tetrahydroquinazolines F. However, we never
observed formamides such as 9 in reactions with other anilines.
Hence, stabilization of 10 and 9 by intramolecular H-bonding with
the neighboring nitro group might account for the unusual course
of this reaction (Scheme 3).
The final interesting point is the condensation of 4-unsub-
stituted anilines, the simplest being aniline itself (4b). Earlier
studies have shown that the reaction between aniline and CH2O in
EtOH/H2O/HCl does not produce the corresponding analogue of
Tro¨ger’s base and leads only to polymeric products.2 However, we
have recently synthesized (ꢀ)-1b in very good yield in CF3COOH.32
Furthermore, it has been shown that 3-halo-2-methyl anilines
produce Tro¨ger’s base analogues in 32–59% yield, though some
polymerization apparently occurs, which possibly accounts for the
modest yields.28 The above-mentioned condensations of 2,5-
dibromoaniline 6a to give up to 67% of (ꢀ)-3a is in line with this
observation. It thus appears clear that at least relatively electron-
poor 4-unsubstituted anilines do give analogues of Tro¨ger’s base
upon condensation with paraformaldehyde in CF3COOH.
We were therefore keen to examine the possibility to perform
this reaction with 4-unsubstituted anilines bearing electron-do-
nating substituents. Gratifyingly, 2-methyl- as well as 2,3-, 2,5-, and
3,5-dimethylanilines 6c–f afforded the corresponding Tro¨ger’s base
derivatives in yields up to 94% (Table 3). However, 2-methoxyani-
line (6g) indeed gave only products of polymerization. One can thus
conclude that an unsubstituted position 4 poses a problem only in
anilines that are strongly activated toward electrophilic sub-
stitution, while moderately activated or non-activated anilines re-
act smoothly. Analogues of Tro¨ger’s base with free 2,8-positions are
interesting substrates for the synthetic approach based on regio-
selective electrophilic substitutions, which was recently introduced
by us.32,44
We believe that seemingly contradictory findings on the be-
havior of 4-unsubstituted anilines in the reaction with formalde-
hyde in various media may be satisfactorily explained as follows.
The first unwanted process that will lead to polymeric products is
the attack of the protonated formaldehyde at the most reactive
para-position of the starting aniline. As discussed above, this pro-
cess is slowed down in CF3COOH compared to HCl/H2O due to
lower concentration of protonated formaldehyde. Another side
reaction that would prevent formation of Tro¨ger’s base is the attack
of the electrophile B at the para-position of the aniline to give para-
aminobenzylamine K. The latter cannot undergo further trans-
formation to give Tro¨ger’s base analogue. The formation of K is
likely even favored over the formation of its isomer ortho-amino-
benzylamine E, for steric reasons. However, the formation of K, as
well as of its isomer E, is reversible (see above), and reverse process
generates back the reactive iminium cation B.41 Thus, the overall
process produces a Tro¨ger’s base analogue (ꢀ)-H, the formation of
which is irreversible. Rather widely varied yields (47–94%) of
Tro¨ger’s base analogues (ꢀ)-3d–f from the corresponding isomeric
dimethylanilines 6d–f might indicate that the para-aminobenzyl-
amine K also undergoes further transformations, that do not lead to
isolable products.
To recapitulate this section, yields of Tro¨ger’s base analogues
from anilines bearing electron-donating substituents were good to
excellent in the standardized conditions (addition of reactants to
CF3COOH at ꢁ15 ꢂC followed by reaction at rt for 24 h). Outcomes of
reactions with anilines bearing electron-withdrawing groups vary
considerably. In some cases, initiating reactions at rt improved
yields, but in no case it provided Tro¨ger’s base analogues that
would otherwise be inaccessible in the standard conditions. Other
variations of reaction conditions, such as elevated temperatures or
prolonged reaction times, had only minor effects. An unsubstituted
para-position in aniline derivatives only presents a problem in
combination with very strong electron-donating groups (e.g., MeO)
in other positions. Anilines with free para-position and moderately
electron-donating or electron-withdrawing groups in other posi-
tions do produce Tro¨ger’s base analogues with yields up to 94%.
¨
2.2. Enantioseparation of Troger’s base analogues
With a number of functionalized Tro¨ger’s base analogues in
hand, we turned to the study of their enantioseparation. In order to
have a more representative library of compounds, we added a few
more functionalized derivatives (ꢀ)-1u–w and (ꢀ)-2h–k that are
unavailableviathecondensationofanilines withformaldehyde. They
were synthesized by alternative methods according to published
procedures (see Fig. 5 for structures and Section 4 for references).
For the separations, we have chosen the commercial ‘brush-
type’ chiral stationary phase (CSP) Whelk O1 with covalently bound
chiral selector derived from 3,4-disubstituted 1,2,3,4-tetrahy-
drophenanthrene (Fig. 6).45 From a practical standpoint, brush-type
CSPs are compatible with a wider range of mobile phases including
highly polar and chlorinated solvents and are more stable than
polysaccharide-coated CSPs, which can leach off with use. This is
important as some analogues of Tro¨ger’s base showed very strong
retention on the cellulose-based CSP Chiralcel OJ, thus rendering
separation rather lengthy.20 Among other brush-type CSPs, Whelk
O1 became popular because of its broad versatility combined with
Figure 5.

6258
D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
Table 5
Enantioseparations of 2,8-disubstituted Tro¨ger’s base analogues
Tro¨ger’s base
analogue
Mobile phase (hexane/
i-PrOH, v/v)
k1
k2
a
Rs
(ꢀ)-1b
(ꢀ)-1a
(ꢀ)-1c
(ꢀ)-1d
(ꢀ)-1e
(ꢀ)-1f
(ꢀ)-1g
(ꢀ)-1h
(ꢀ)-1j
(ꢀ)-1l
(ꢀ)-1m
(ꢀ)-1n
(ꢀ)-1o
(ꢀ)-1p
(ꢀ)-1q
(ꢀ)-1r
(ꢀ)-1s
(ꢀ)-1u
(ꢀ)-1v
(ꢀ)-1w
95:5
95:5
95:5
95:5
95:5
95:5
95:5
95:5
80:20
95:5
95:5
95:5
95:5
95:5
95:5
50:50
80:20
50:50
80:20
95:5
1.33
2.17
1.63
2.34
2.46
2.49
3.41
3.65
1.51
1.79
1.30
1.32
1.49
1.54
1.62
1.07
1.45
1.24
1.00
1.00
1.00
2.05
2.23
4.57
3.51
3.67
3.82
4.16
2.21
4.07
1.33
1.21
2.08
2.60
3.02
0.55
1.01
0.72
N.S.a
N.S.a
N.S.a
5.79
1.35
0.99
0.74
0.41
0.53
6.90
5.06
5.67
1.10
1.09
1.37
1.56
22.6
0.49
8.46
12.0
8.5
3.16
2.42
1.83
1.41
1.95
10.5
Figure 6. (3R,4S)-Whelk O1 CSP used in this study. This CSP is marketed by Regis
Technologies, Inc under the name (S,S)-Whelk O1. The absolute configuration of the
chiral selector is thus incorrectly designated. This should not lead to confusion. In-
correct designation results from the fact that the original version of this CSP had an 11-
carbon linker, but substitution of it by the 3-carbon linker results in the inversion of
Cahn–Ingold–Prelog priorities.
9.07
7.40
1.46
1.62
2.11
2.53
24.1
0.70
10.5
excellent efficiency. A considerable amount of data on enantiose-
parations of various analytes have been accumulated and ratio-
nalized.46 Importantly, a small well-defined chiral selector appears
beneficial for the understanding of structure–enantioselectivity
relationship.
Due to conformational preferences of the saturated ring in
1,2,3,4-tetrahydrophenanthrene (half-chair with the pseudoaxial
amide group), the chiral selector of Whelk O1 has a cleft-like shape.
Preferential binding of the more retained enantiomer of a chiral
analyte in the cleft is provided through simultaneous face-to-face
13.5
3.87
7.91
a
N.S.¼no separation.
two effects is clearly illustrated by the comparison between (ꢀ)-1g
and (ꢀ)-1h substituted with electron-donating MeO and MeS
groups, respectively. In the former case, retention times of both
enantiomers are longer, the separation factor is smaller, the reso-
lution is poorer, and the tailing of peaks is considerabledall this
because of the stronger interactions of the more polar MeO group
with the underlying SiO2 support. Hydrogen bond donors
(CH2CH2OH, NH2, (ꢀ)-1j,u) at the para-position strongly increase
retention and decrease enantioselectivity due to pronounced in-
teractions with SiO2.
Alkyl substituents increase the separation factors (cf. (ꢀ)-1a,c–f
vs (ꢀ)-1b). Moreover, both the increase of length of alkyl sub-
stituents (from Me to hexyl) and the increase of their volume (from
Et to t-Bu) progressively decreases retention of both enantiomers,
but at the same time enhances enantioselectivity. The same trend
has been observed previously for 3-alkyl-5-methyl-5-arylhy-
dantoins bearing alkyl substituents of various lengths.49 Phenyl
group in the para-position improves the enantioselectivity com-
pared to the unsubstituted parent molecule (cf. (ꢀ)-1w vs (ꢀ)-1b)
and only slightly diminishes it compared to the methyl group (cf.
(ꢀ)-1w vs (ꢀ)-1a). One can conclude that a bulky substituent does
p–
p
interactions with the
p-acidic 3,5-dinitrobenzoyl moiety, face-
to-edge binding with the p
p
–
p
-basic naphthalene system and H-
bonding with the hydrogen of the amide group. The less retained
enantiomer is incapable of all these interactions without inducing
a deviation from the lowest-energy conformation. This model is
supported by systematic separations of various analytes, as well as
by crystallization of homo- and heterochiral complexes of the chiral
selector and an analyte, by analysis of 1H NMR spectra of homo- and
heterochiral complexes, and by computational studies.47 Thus,
a typical good analyte for Whelk O1 CSP is an aromatic system
with an additional H-bond acceptor in the proximity of the chiral
center.
Since the tertiary nitrogen atoms in the molecule of Tro¨ger’s
base are asymmetric centers and can act as H-bond acceptors,
Tro¨ger’s base certainly appears a promising analyte for separation
on Whelk O1. This is confirmed by the efficient separation of
Tro¨ger’s base on Whelk O1 reported by the manufacturer (k1¼2.52,
a
¼1.80 in hexane/EtOH 96:4).48 In order to gain insight into the
structure–enantioselectivity relationship, we proceeded with the
separation of variously functionalized analogues of Tro¨ger’s base.
Experimental data are summarized in Tables 5 and 6 as retention
Table 6
factors of the two enantiomers k1, k2, separation factors
a
¼k2/k1,
Enantioseparations of Tro¨ger’s base analogues bearing substituents in positions
other than 2 and 8
and base line resolutions Rs. As long as it was practical, we used
hexane/i-PrOH (95:5) as a mobile phase, for the purpose of better
direct comparison between different analytes.
Tro¨ger’s base
analogue
Mobile phase
(hexane/i-PrOH, v/v)
k1
k2
a
Rs
From the experiments on the enantioseparation of derivatives
(ꢀ)-1a–w (bearing substituents only in the para-position relative to
the nitrogen), a few trends became apparent. As expected, the
electron-withdrawing substituents reduce separation factors and
increase retention times. In this case two cooperative factors ruin
(ꢀ)-2a
(ꢀ)-2b
(ꢀ)-2c
(ꢀ)-2d
(ꢀ)-2e
(ꢀ)-2g
(ꢀ)-2h
(ꢀ)-2i
(ꢀ)-2j
(ꢀ)-2k
(ꢀ)-3a
(ꢀ)-3b
(ꢀ)-3c
(ꢀ)-3d
(ꢀ)-3e
(ꢀ)-3f
98:2
98:2
98:2
95:5
90:10
98:2
95:5
95:5
80:20
80:20
98:2
100:0
99:1
98:2
95:5
98:2
0.49
0.61
1.24
1.31
1.00
1.07
1.00
1.21
1.14
1.70
1.00
1.35
1.14
1.00
1.00
1.20
1.00
1.41
0.53
1.79
N.S.a
0.32
N.S.a
1.43
0.88
3.77
N.S.a
2.31
0.89
N.S.a
N.S.a
0.58
N.S.a
2.12
1.18
0.58
2.95
10.4
1.43
14.6
4.88
6.51
3.93
1.05
1.41
0.72
0.75
1.42
0.69
1.55
3.17
1.73
the separation: firstly, the face-to-face
p–p binding between the
16.6
8.28
electron-poor aromatic ring and the electron-poor part of the chiral
selector is not favored, and secondly, the ‘achiral retention’ due to
dipole–dipole interactions with the underlying silica support is
increased. In case of electron-donating substituents in the analytes,
these effects become competitive: the favored binding of the
5.30
1.21
analyte with the
p-acidic 3,5-dinitrobenzoyl moiety of the chiral
0.90
0.97
selector is beneficial for enantioselectivity, but at the same time
a polar substituent increases achiral retention and has a negative
impact on the enantioselectivity. Subtle interplay between these
a
N.S.¼no separation.

D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
6259
not alter the conformational preference of the analyte, while achiral
interactions with underlying silica support are reduced.
derivative of Tro¨ger’s base will be feasible with the aid of the de-
scribed method.
For the series of halogen derivatives (ꢀ)-1l–o, there is a constant
increase in enantioselectivity with the increase of the polarizability
of the halogen (from F to I). This is in perfect agreement with the
observations of Pirkle and co-workers on the separations of amide
derivatives of 1-(4-halophenyl)ethylamine on Whelk O1. Our
finding also supports the generalized prediction of Pirkle and co-
workers that polarizable halogens will increase enantioselectivity,
Perspectives of the reported study are numerous and include:
(1) further variations of the structure of Tro¨ger’s base analogues, in
order to prove predictive capacity of the method; (2) getting an
insight into the quantitative structure–enantioselectivity relation-
ship through a combination of various techniques: further chro-
matographic data, including correlations between elution order
and absolute configurations of analytes, 1H NMR and crystallo-
graphic studies of homo- and heterochiral diasteroisomeric com-
plexes of analytes and the chiral selector, and computer modeling
may prove useful; (3) expanding the scope of separation through
the use of other commercial CSPs. All these studies are currently in
progress and their results will be reported in due course.
as long as the halogen-substituted aromatic ring is p-basic and the
inductive effect of the substituent does not affect the strength of
the hydrogen bond involved in the chiral recognition.50 The fact
that derivatives with Br or I in different positions ((ꢀ)-1n,o,
(ꢀ)-2b,g) can be satisfactorily separated is rather valuable, in view
of possibilities to carry out various cross-coupling reactions with
these intermediates.38,39 When performed in non-acidic condi-
tions, transformations of these derivatives are unlikely to be ac-
companied by racemization.
4. Experimental section
4.1. General
In the series of derivatives (ꢀ)-2 and (ꢀ)-3 bearing substituents
in positions other than para relative to the nitrogen, there is
a general drastic decrease of separation factors compared to the
unsubstituted parent molecule (ꢀ)-1b, regardless of the nature of
the substituents (Table 6). In agreement with the rationalization
given above, both strong electron-donating and electron-with-
drawing groups were particularly detrimental for enantiose-
lectivity. Notable exceptions are derivatives with H-bond donor
groups in ortho-position respective to the nitrogen atom, such as
(ꢀ)-2i,k, which demonstrated relatively good separation. This
suggests additional hydrogen bonding, presumably, with NO2
groups of the chiral selector as hydrogen bond acceptors.
All chemicals were purchased from Aldrich, Acros or TCI Europe
and used without further purification unless stated otherwise.
Column chromatography: SiO2 Kieselgel 60 (Macherey-Nagel, par-
ticle size 0.04–0.063 mm). TLC: precoated SiO2 plates Kieselgel
60F254 (Merck). 1H NMR (300 MHz) and 13C NMR (75 MHz) spectra
were recorded on a Brucker Avance 300 spectrometer; chemical
shifts (d) are given in parts per million relative to Me4Si; coupling
constants (J) are given in hertz. Electron impact mass spectra
(EIMS) were recorded on a Waters AutoSpec 6F instrument and
electrospray ionization mass spectra (ESIMS) on a Waters QToF 2
instrument; m/z with the lowest isotopic mass are reported. The
following derivatives were synthesized according to published
methods: (ꢀ)-1s,u,39 (ꢀ)-1v,6 (ꢀ)-1w,51 (ꢀ)-2b,20 (ꢀ)-2h,i,39
(ꢀ)-2j,k.20
Remarkably, all di- and tetramethyl derivatives (ꢀ)-2a,3c–f with
substituents in positions other than para respective to the nitrogen
showed smaller separation factors than Tro¨ger’s base (ꢀ)-1a. This is
clearly indicative of the steric hindrance, which prevents efficient
HPLC separations were performed at ambient temperature on
an Agilent 1100 instrument equipped with the Rheodyne 7725
manual injector. The solvents were of HPLC grade. Column (S,S)-
Whelk-O1 (250ꢃ4.6 mm) was purchased from Regis Technologies
p–p interactions between the aromatic rings of Tro¨ger’s base
framework and the chiral selector. Finally, octafluoride (ꢀ)-3b
predictably showed very poor retention even with hexane as
a mobile phase and no enantioseparation was achieved.
Overall, it appears that 2,8-disubstituted Tro¨ger’s base ana-
logues are the easiest to separate by the described method, while
4,10-positions (ortho to the nitrogen atoms) should either remain
unsubstituted or bear a hydrogen bond donor. This consideration
should probably be kept in mind in the design of studies that rely
on enantiomerically pure Tro¨ger’s base analogues.
(USA). Injection: 20
m
L of analyte solution in CH2Cl2, ca. 1 mg mLꢁ1
;
;
mobile phase: hexane/i-PrOH, nominal flow rate: 2.0 mL minꢁ1
detection: UV at fixed wavelength 254 or 230 nm. Void volume was
detected from the injection of 1,3,5-tri-tert-butylbenzene. Separa-
tion parameters were calculated as follows: k1¼(t1ꢁt0)/t0,
k2¼(t2ꢁt0)/t0,
a
¼k2/k1, Rs¼2(t2ꢁt1)/(w2þw1), where t1, t2 are re-
tention times of the two enantiomers, t0 is the void time, k1, k2 are
retention factors of the two enantiomers, is the separation factor,
a
3. Conclusion
w1, w2 are widths of peaks at the base line, Rs is the resolution at the
base line.
Firstly, we studied the scope and limitations of a general pro-
cedure for the synthesis of Tro¨ger’s base analogues from anilines
and paraformaldehyde in CF3COOH. It appears to be a method of
choice for all anilines since it provides better yields compared to
any other known protocol. The scope is broad and the reaction
conditions are identical for most anilines, though in a few cases
minor modifications of the general procedure may be useful. The
notable limitations are the presence of very strong electron-do-
nating groups (Me2N) and very strong acceptors, such as NO2 and
CN, when not counterbalanced with electron-donating groups. A
number of Tro¨ger’s base analogues prepared in the course of the
present study can serve as useful intermediates for various syn-
thetic transformations.
Secondly, we systematically studied separations of Tro¨ger’s base
analogues on the commercially available CSP Whelk O1. Structure–
enantioselectivity relationships were established in terms of the
substitution pattern of the aromatic rings and the nature of sub-
stituents. This mechanistic explanation can be used to predict
whether or not enantioseparation of a given, perhaps yet unknown
Crystallographic data (excluding structure factors) for the
structures in this paper have been deposited with the Cambridge
Crystallographic Data Centre as supplementary publication no.
CCDC 681274 ((ꢀ)-3a), CCDC 685743 (7), and CCDC 682620 (9).
Copies of the data can be obtained, free of charge, on application to
CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (fax: þ44 (0)12233
36033 or e-mail: deposit@ccdc.cam.ac.uk).
¨
4.2. Synthesis of Troger’s base analogues ( )-1–3 from
anilines 4–6
Anilines 4–6 (5 mmol) and then paraformaldehyde (300 mg,
10 mmol) were added in portions under vigorous stirring to
CF3COOH (10 mL) at ꢁ15 ꢂC. The resulting mixture was allowed to
reach rt and stirred for 24 h, then slowly added to a stirred mixture
of ice and 30% aqueous NH3 (17 mL). If necessary, pH value of the
resulting mixture was adjusted to 9–10 by the addition of 30%
aqueous NH3. Extraction with CH2Cl2 (3ꢃ50 mL), drying of the or-
ganic layer over MgSO4, and removal of the solvent in vacuum gave

6260
D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
a crude product, which was purified by column chromatography on
SiO2 (details are specified below for individual compounds). In
some cases reaction conditions were varied as specified in Tables 1–
3, but the work up and purification procedures were always the
same for a given compound.
For the following Tro¨ger’s base analogues spectral data were
identical with those published before: (ꢀ)-1a,24 (ꢀ)-1b,32 (ꢀ)-1j,24
(ꢀ)-1k,6 (ꢀ)-1l–o,27 (ꢀ)-1p,4 (ꢀ)-1q,25 (ꢀ)-2a,d,24 (ꢀ)-2e,9 and
(ꢀ)-3b.35 The following compounds have been reported before, but
without spectral characterization: (ꢀ)-1g,h,25 and (ꢀ)-2g.34 We
therefore give below their 1H and 13C spectra. For all other Tro¨ger’s
base analogues we report full analytical details below.
29.1, 31.4, 31.7, 35.4, 58.6, 66.9,124.8,126.5,127.3, 127.5, 138.6,145.7.
HREIMS: m/z: calcd for
390.3025.
C
27H38N2 ([M]þ) 390.3035; found:
4.2.5. (ꢀ)-2,8-Dimethoxy-6H,12H-5,11-methanodibenzo[b,f][1,5]-
diazocine ((ꢀ)-1g)
Prepared according to the general procedure from 4-methoxy-
aniline 4g (615 mg, 5.00 mmol). The crude product was purified by
column chromatography (CH2Cl2/AcOEt 4:1) to afford pure (ꢀ)-1g
(469 mg, 66%) as a white solid; mp 171–173 ꢂC (lit.26 mp 172–
173 ꢂC). 1H NMR (300 MHz, CDCl3, 25 ꢂC):
d
¼3.70 (s, 6H), 4.08 (d,
2JH,H¼17.8 Hz, 2H), 4.29 (s, 2H), 4.65 (d, 2JH,H¼17.8 Hz, 2H), 6.42 (d,
4JH,H¼3.0 Hz, 2H), 6.74 (dd, 3JH,H¼8.7 Hz, 4JH,H¼3.0 Hz, 2H), 7.06 (d,
13
4.2.1. (ꢀ)-2,8-Diethyl-6H,12H-5,11-methanodibenzo[b,f][1,5]-
diazocine ((ꢀ)-1c)
3JH,H¼8.7 Hz, 2H). C NMR (75 MHz, CDCl3, 25 ꢂC):
d¼55.3, 58.9,
67.2, 110.9, 114.0, 125.9, 128.6, 140.9, 156.1.
Prepared according to the general procedure from 4-ethylani-
line 4c (605 mg, 5.00 mmol). The crude product was purified by
column chromatography (CH2Cl2/AcOEt 5:1) to afford pure (ꢀ)-1c
(597 mg, 86%) as a colorless oil, which partially solidified upon
standing for a few weeks, but did not produce a material with
4.2.6. (ꢀ)-2,8-Bis(methylsulfanyl)-6H,12H-5,11-methanodibenzo-
[b,f][1,5]diazocine ((ꢀ)-1h)
Prepared according to the general procedure from amine 4h
(695 mg, 5.00 mmol). The crude product was purified by column
chromatography (CH2Cl2/AcOEt 4:1, then 1:1) to afford pure (ꢀ)-1h
(527 mg, 69%) as an off-white solid; mp 186–187 ꢂC (lit.26 mp 184–
a
distinct mp. 1H NMR (300 MHz, CDCl3, 25 ꢂC):
d
¼1.15 (t,
3
2
3JH,H¼7.7 Hz, 6H), 2.51 (q, JH,H¼7.7 Hz, 4H), 4.13 (d, JH,H¼16.7 Hz,
2
4
2H), 4.29 (s, 2H), 4.66 (d, JH,H¼16.7 Hz, 2H), 6.72 (d, JH,H¼1.8 Hz,
185 ꢂC). 1H NMR (300 MHz, CDCl3, 25 ꢂC):
d
¼2.40 (s, 6H), 4.10 (d,
3
4
3
2H), 6.98 (dd, JH,H¼8.1 Hz, JH,H¼1.8 Hz, 2H), 7.05 (d, JH,H¼8.1 Hz,
2JH,H¼16.9 Hz, 2H), 4.28 (s, 2H), 4.64 (d, 2JH,H¼16.9 Hz, 2H), 6.82 (d,
2H). 13C NMR (75 MHz, CDCl3, 25 ꢂC):
d
¼15.4, 28.2, 58.6, 66.9, 124.9,
4JH,H¼2.2 Hz, 2H), 7.05 (d, 3JH,H¼8.4 Hz, 2H), 7.10 (dd, JH,H¼8.4 Hz,
3
13
126.0, 126.8, 127.6, 139.7, 145.7. HREIMS: m/z: calcd for C19H22N2
4JH,H¼2.2 Hz, 2H). C NMR (75 MHz, CDCl3, 25 ꢂC):
d¼16.6, 58.6,
([M]þ) 278.1783; found: 278.1788.
66.9, 125.5, 125.7, 126.8, 128.4, 133.1, 145.6. HREIMS: m/z: calcd for
C
17H18N2S2 ([M]þ) 314.0911; found: 314.0908.
4.2.2. (ꢀ)-2,8-Di(iso-propyl)-6H,12H-5,11-methanodibenzo[b,f]-
[1,5]diazocine ((ꢀ)-1d)
4.2.7. (ꢀ)-[(6H,12H-5,11-Methanodibenzo[b,f][1,5]diazocin-2,8-
diyl)bis(4,1-phenylene)]dicarbonitrile ((ꢀ)-1r)
Prepared according to the general procedure from 4-(iso-pro-
pyl)aniline 4d (675 mg, 5.00 mmol). The crude product was puri-
fied by column chromatography (CH2Cl2/AcOEt 5:1) to afford pure
Prepared according to the general procedure from amine 4r
(487 mg, 2.50 mmol), paraformaldehyde (150 mg, 5.00 mmol), and
CF3COOH (5 mL). The crude product was purified by column
chromatography (CH2Cl2/AcOEt 9:1) to afford pure (ꢀ)-1r (0.231 g,
(ꢀ)-1d (655 mg, 86%) as a white solid; mp 109–110.5 ꢂC. 1H NMR
3
(300 MHz, CDCl3, 25 ꢂC):
d
2
¼1.16 (d, JH,H¼6.9 Hz, 12H), 2.78 (sept,
3JH,H¼6.9 Hz, 2H), 4.14 (d, JH,H¼16.7 Hz, 2H), 4.29 (s, 2H), 4.68 (d,
43%) as a white solid; mp 109–110.5 ꢂC. 1H NMR (300 MHz, CDCl3,
2JH,H¼16.7 Hz, 2H), 6.74 (d, 4JH,H¼1.8 Hz, 2H), 7.02 (dd, 3JH,H¼8.4 Hz,
25 ꢂC):
d
¼4.30 (d, JH,H¼16.8 Hz, 2H), 4.38 (s, 2H), 4.81 (d,
2
4JH,H¼1.8 Hz, 2H), 7.07 (d, JH,H¼8.4 Hz, 2H). 13C NMR (75 MHz,
2JH,H¼16.8 Hz, 2H), 7.17 (d, JH,H¼2.1 Hz, 2H), 7.26 (d, JH,H¼8.4 Hz,
3
4
3
4
3
CDCl3, 25 ꢂC):
d
¼23.9, 33.5, 58.5, 66.8, 124.5, 124.9, 125.4, 127.5,
2H), 7.42 (dd, 3JH,H¼8.4 Hz, JH,H¼2.1 Hz, 2H), 7.57 (d, JH,H¼8.7 Hz,
144.3, 145.8. HREIMS: m/z: calcd for C21H26N2 ([M]þ) 306.2096;
4H), 7.66 (d, JH,H¼8.7 Hz, 4H). C NMR (75 MHz, CDCl3, 25 ꢂC):
3
13
found: 306.2091.
d
¼58.8, 66.9, 110.6, 118.9, 125.8, 125.8, 126.4, 127.3, 128.4, 132.6,
135.0, 145.0, 148.7; HRESIMS: m/z: calcd for C29H20N4Na ([MþNa]þ)
4.2.3. (ꢀ)-2,8-Di(tert-butyl)-6H,12H-5,11-methanodibenzo[b,f]-
[1,5]diazocine ((ꢀ)-1e)
447.1586; found: 447.1581.
Prepared according to the general procedure from 4-(tert-
butyl)aniline 4e (745 mg, 5.00 mmol). The crude product was pu-
rified by column chromatography (CH2Cl2/AcOEt 5:1) to afford pure
(ꢀ)-1e (733 mg, 88%) as a white solid; mp 207–209 ꢂC. 1H NMR
4.2.8. (ꢀ)-4,10-Dibromo-2,8-bis(trifluoromethyl)-6H,12H-5,11-
methanodibenzo[b,f][1,5]diazocine ((ꢀ)-2c)
Prepared according to the general procedure from amine 5c
(502 mg, 2.1 mmol), paraformaldehyde (130 mg, 4.30 mmol), and
CF3COOH (4 mL). The crude product was purified by column
chromatography (hexane/CH2Cl2 85:15) to afford pure (ꢀ)-2c
(300 MHz, CDCl3, 25 ꢂC):
d
¼1.24 (s, 18H), 4.15 (d, 2JH,H¼16.8 Hz, 2H),
2
4
4.28 (s, 2H), 4.69 (d, JH,H¼16.8 Hz, 2H), 6.90 (d, JH,H¼1.8 Hz, 2H),
3
3
4
7.07 (d, JH,H¼8.4 Hz, 2H), 7.18 (dd, JH,H¼8.4 Hz, JH,H¼1.8 Hz, 2H).
(0.181 g, 34%) as a white solid; mp 115–116 ꢂC. 1H NMR (300 MHz,
13C NMR (75 MHz, CDCl3, 25 ꢂC):
d
¼31.3, 34.2, 58.5, 66.8, 123.4,
CDCl3, 25 ꢂC):
d
¼4.37 (s, 2H), 4.43 (d, JH,H¼17.6 Hz, 2H), 4.67 (d,
2
4
4
124.3, 124.6, 127.1, 145.6, 146.6. HREIMS: m/z: calcd for C23H30N2
2JH,H¼17.6 Hz, 2H), 7.24 (d, JH,H¼1.1 Hz, 2H), 7.73 (d, JH,H¼1.1 Hz,
([M]þ) 334.2409; found: 334.2404.
2H). 13C NMR (75 MHz, CDCl3, 25 ꢂC):
d
¼55.2, 67.2, 120.4, 123.0 (q,
2
1JC,F¼272.4 Hz, CF3), 123.5 (q, JC,F¼3.6 Hz), 127.9 (q, JC,F¼33.1 Hz,
3
3
4.2.4. (ꢀ)-2,8-Dihexyl-6H,12H-5,11-methanodibenzo[b,f][1,5]-
diazocine ((ꢀ)-1f)
CCF3), 127.9 (q, JC,F¼3.6 Hz), 130.9, 147.8. HRESIMS: m/z: calcd for
79
C
H
Br2F6N2 ([MþH]þ) 514.9193; found: 514.9209.
17 11
Prepared according to the general procedure from 4-hexylani-
line 4f (885 mg, 5.00 mmol). The crude product was purified by
column chromatography (CH2Cl2/AcOEt 5:1) to afford pure (ꢀ)-1f
(770 mg, 79%) as a white solid; mp 61–63 ꢂC. 1H NMR (300 MHz,
4.2.9. (ꢀ)-2,4,8,10-Tetrabromo-6H,12H-5,11-methanodibenzo[b,f]-
[1,5]diazocine ((ꢀ)-2g)
Prepared according to the general procedure from amine 5g
(1.255 g, 5 mmol). The crude product was purified by column
chromatography (CH2Cl2) to afford pure (ꢀ)-2g (269 mg, 20%) to-
gether with 6,8-dibromo-3-(2,4-dibromophenyl)-1,2,3,4-tetrahy-
droquinazoline (7, 671 mg, 51%) and N-methyl-2,4-dibromoaniline
(8, 60 mg, 4.5%).
CDCl3, 25 ꢂC):
d
¼0.87 (t, 3JH,H¼6.7 Hz, 6H), 1.20–1.35 (m, 12H), 1.52
3
3
(quint, JH,H¼6.7 Hz, 4H), 2.46 (t, JH,H¼7.8 Hz, 4H), 4.12 (d,
2JH,H¼16.8 Hz, 2H), 4.29 (s, 2H), 4.66 (d, JH,H¼16.8 Hz, 2H), 6.70
2
(d, 4JH,H¼1.8 Hz, 2H), 6.96 (dd, 3JH,H¼8.2 Hz, 4JH,H¼1.8 Hz, 2H), 7.03
(d, 3JH,H¼8.2 Hz, 2H). 13C NMR (75 MHz, CDCl3, 25 ꢂC):
¼14.1, 22.6,
d

D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
6261
Analytical data of (ꢀ)-2g: white solid, mp 222–224 ꢂC. 1H NMR
(ꢀ)-3f (326 mg, 47%) as a white solid; mp 195–197 ꢂC. 1H NMR
2
(300 MHz, CDCl3, 25 ꢂC):
d
¼4.23–4.33 (m, 4H), 4.55 (d, JH,H
¼
(300 MHz, CDCl3, 25 ꢂC):
d¼2.06 (s, 6H), 2.40 (s, 6H), 3.93 (d,
17.8 Hz, 2H), 7.09 (d, 4JH,H¼2.1 Hz, 2H), 7.59 (d, 4JH,H¼2.1 Hz, 2H). 13
C
2JH,H¼16.9 Hz, 2H), 4.27 (s, 2H), 4.41 (d, 2JH,H¼16.9 Hz, 2H), 6.76 (d,
3
NMR (75 MHz, CDCl3, 25 ꢂC):
d
¼55.0, 67.4, 117.7, 120.7, 129.1, 132.1,
3JH,H¼7.5 Hz, 2H), 6.97 (d, JH,H¼7.5 Hz, 2H). 13C NMR (75 MHz,
79
134.0, 143.6. ESIMS: m/z: 535 ([C
H
15 11
Br4N2]þ, [MþH]þ).
CDCl3, 25 ꢂC):
d
¼17.0, 17.8, 53.6, 66.3, 125.1, 126.5, 128.5, 130.3,
Analytical data of 7: off-white solid, mp 155–157 ꢂC. 1H NMR
132.8, 146.4; HREIMS: m/z: calcd for C19H22N2 ([M]þ) 278.1783;
(300 MHz, CDCl3, 25 ꢂC):
d
¼4.39 (s, 2H, ArCH2N), 4.51 (br, 1H, NH),
found: 278.1783.
4.67 (br, 2H, NCH2N), 6.90 (d, 3JH,H¼8.6 Hz,1H), 7.07 (d, 4JH,H¼2.2 Hz,
3
4
4
1H), 7.27 (dd, JH,H¼8.6 Hz, JH,H¼2.3 Hz, 1H), 7.45 (d, JH,H¼2.2 Hz,
4.2.15. N-[4-Methyl-2-[(4-methyl-2-nitro-phenylamino)methyl]-6-
nitrophenyl]-formamide (9)
4
13
1H), 7.72 (d, JH,H¼2.3 Hz, 1H). C NMR (75 MHz, CDCl3, 25 ꢂC):
d
¼52.1, 61.9,109.6,109.9,117.0,120.3,122.8,123.9,128.3,131.3,132.8,
Reaction between 4-methyl-2-nitroaniline 5f and para-
formaldehyde was carried out according to the general procedure
with the following amendment: after addition of the reactants to
CF3COOH, the reaction mixture was heated to reflux for 24 h. Col-
umn chromatography (CH2Cl2/AcOEt 8:2) afforded 9 (337 mg, 39%)
as an orange solid; mp 192–193 ꢂC. 1H NMR (300 MHz, CDCl3,
79
136.0, 139.7, 147.0. EIMS: m/z: 522 ([C
H
Br4N2]þ, [M]þ).
14 10
Analytical data of 8 were identical to those published earlier.52
4.2.10. (ꢀ)-1,4,7,10-Tetrabromo-6H,12H-5,11-methanodibenzo-
[b,f][1,5]diazocine ((ꢀ)-3a)
Prepared according to the general procedure from amine 6a
(1.255 g, 5 mmol). The crude product was purified by column
chromatography (CH2Cl2) to afford pure (ꢀ)-3a (820 mg, 61%) as
25 ꢂC):
d
¼2.27 (s, 3H), 2.38 (s, 3H), 4.54 (d, 3JH,H¼6.0 Hz, 2H, CH2N),
6.59 (d, 3JH,H¼8.6 Hz, 1H), 7.22 (dd, 3JH,H¼8.6 Hz, 4JH,H¼2.0 Hz, 1H),
4
7.49 (s, 1H), 7.81 (s, 1H), 8.02 (d, JH,H¼2.0 Hz, 1H), 8.35 (t,
1
a white solid; mp 232–234 ꢂC. H NMR (300 MHz, CDCl3, 25 ꢂC):
3JH,H¼6.0 Hz, 1H, CH2NH), 8.44 (s, 1H, CH]O), 8.66 (br, 1H,
2
2
d
¼4.27 (s, 2H), 4.34 (d, JH,H¼17.8 Hz, 2H), 4.41 (d, JH,H¼17.8 Hz,
NHCH]O). 13C NMR (75 MHz, (CD3)2CO, 25 ꢂC; assignments from
3 3
2H), 7.18 (d, JH,H¼8.6 Hz, 2H), 7.37 (d, JH,H¼8.6 Hz, 2H). 13C NMR
DEPT experiments):
d
¼20.8 (CH3), 21.8 (CH3), 44.9 (CH2), 116.5 (CH),
(75 MHz, CDCl3, 25 ꢂC):
d
¼56.9, 66.6, 119.5, 122.1, 129.5, 130.2,
125.7 (CH), 126.7 (C), 127.3 (C), 127.5 (CH), 134.2 (CH), 139.4 (CH),
139.7 (C), 140.0 (C), 144.9 (C), 145.0 (C), 148.6 (C), 161.9 (CH). HRE-
SIMS: m/z: calcd for C16H16N4NaO5 ([MþNa]þ) 367.1018; found:
367.1015.
79
132.4, 146.5. HRESIMS: m/z: calcd for C
534.7656; found: 534.7657.
H
15 11
Br4N2 ([MþH]þ)
4.2.11. (ꢀ)-4,10-Dimethyl-6H,12H-5,11-methanodibenzo[b,f]-
[1,5]diazocine ((ꢀ)-3c)
Acknowledgements
Prepared according to the general procedure from o-toluidine 6c
(536 mg, 5.00 mmol). The crude product was purified by column
chromatography (hexane/AcOEt 95:05) to afford pure (ꢀ)-3c
(474 mg, 75%) as a white solid; mp 96–97 ꢂC (lit.16 mp 96–98 ꢂC). 1H
NMR (300 MHz, CDCl3, 25 ꢂC): nearly identical to published data of
(ꢀ)-3c prepared by reductive dehalogenation of the corresponding
S.S. is indebted to Prof. Y. Geerts for the opportunity to conduct
an independent research program in his laboratory and for gener-
ous financial support. We thank Prof. M. Luhmer and R. d’Orazio for
the measurement of NMR spectra and A. Remacle and S. M’Zoudi
for the technical assistance. This study was in part supported by
FNRS (grant 1.5.110.08). B.T. is a FRIA fellow. A.C. thanks IWT-
Vlaanderen for a Ph.D. grant.
dibromo derivative.16 13C NMR (75 MHz, CDCl3, 25 ꢂC):
68.3, 124.6, 125.3, 128.9, 129.7, 133.8, 147.0.
d¼18.0, 55.9,
4.2.12. (ꢀ)-3,4,9,10-Tetramethyl-6H,12H-5,11-methanodibenzo-
[b,f][1,5]diazocine ((ꢀ)-3d)
Supplementary data
Prepared according to the general procedure from 2,3-dime-
thylaniline 6d (606 mg, 5.00 mmol). The crude product was puri-
fied by column chromatography (CH2Cl2) to afford pure (ꢀ)-3d
Additional X-ray diffraction data, including experimental
details, for (ꢀ)-3a, 7, and 9. Supplementary data associated with
this article can be found in the online version, at doi:10.1016/
j.tet.2008.04.111.
(653 mg, 94%) as a white solid; mp 157–159 ꢂC. 1H NMR (300 MHz,
2
CDCl3, 25 ꢂC):
d
¼2.22 (s, 6H), 2.31 (s, 6H), 3.91 (d, JH,H¼16.5 Hz,
2
3
2H), 4.33 (s, 2H), 4.58 (d, JH,H¼16.5 Hz, 2H), 6.66 (d, JH,H¼7.6 Hz,
3
13
2H), 6.82 (d, JH,H¼7.6 Hz, 2H). C NMR (75 MHz, CDCl3, 25 ꢂC):
References and notes
d
¼13.2, 20.2, 55.4, 67.6, 123.8, 125.3 (2C), 131.1, 135.8, 145.8. HRE-
SIMS: m/z: calcd for C19H23N2 ([MþH]þ) 279.1861; found: 279.1855.
1. Valı´k, M.; Strongin, R. M.; Kra´l, V. Supramol. Chem. 2005, 17, 347–367.
2. Dolensky, B.; Elguero, J.; Kral, V.; Pardo, C.; Valık, M. Adv. Heterocycl. Chem.
2007, 93, 1–56.
´
´
4.2.13. (ꢀ)-1,3,7,9-Tetramethyl-6H,12H-5,11-methanodibenzo-
[b,f][1,5]diazocine ((ꢀ)-3e)
3. Adrian, J. C., Jr.; Wilcox, C. S. J. Am. Chem. Soc. 1989, 111, 8055–8057; Adrian, J. C.,
Jr.; Wilcox, C. S. J. Am. Chem. Soc. 1992, 114, 1398–1403.
4. Goswami, S.; Ghosh, K.; Dasgupta, S. J. Org. Chem. 2000, 65, 1907–1914.
5. Paliwal, S.; Geib, S.; Wilcox, C. S. J. Am. Chem. Soc. 1994, 116, 4497–4498;
Bhayana, B.; Wilcox, C. S. Angew. Chem., Int. Ed. 2007, 46, 6833–6836; Fischer, F.
R.; Schweizer, W. B.; Diederich, F. Angew. Chem., Int. Ed. 2007, 46, 8270–8273;
Hof, F.; Scofield, D. M.; Schweizer, W. B.; Diederich, F. Angew. Chem., Int. Ed.
2004, 43, 5056–5059.
Prepared according to the general procedure from 3,5-dime-
thylaniline 6e (606 mg, 5.00 mmol). The crude product was puri-
fied by column chromatography (CH2Cl2/AcOEt 5:1) to afford pure
(ꢀ)-3e (492 mg, 71%) as a white solid; mp 202–203 ꢂC. 1H NMR
(300 MHz, CDCl3, 25 ꢂC):
d
¼2.06 (s, 6H), 2.24 (s, 6H), 4.10 (d,
6. Sergeyev, S.; Scha¨r, M.; Seiler, P.; Lukoyanova, O.; Echegoyen, L.; Diederich, F.
Chem.dEur. J. 2005, 11, 2284–2294.
2JH,H¼16.7 Hz, 2H), 4.24 (s, 2H), 4.49 (d, 2JH,H¼16.7 Hz, 2H), 6.65 (s,
7. Wong, W. W. H.; Diederich, F. Chem.dEur. J. 2006, 12, 3463–3471.
8. Sergeyev, S.; Diederich, F. Angew. Chem., Int. Ed. 2004, 43, 1738–1740.
9. Kiehne, U.; Weilandt, T.; Lu¨tzen, A. Org. Lett. 2007, 9, 1283–1286.
10. Kiehne, U.; Lu¨tzen, A. Eur. J. Org. Chem. 2007, 5703–5711; Kiehne, U.; Weilandt,
T.; Lu¨tzen, A. Eur. J. Org. Chem. 2008, 2056–2064.
2H), 6.82 (s, 2H). 13C NMR (75 MHz, CDCl3, 25 ꢂC):
66.0, 123.18, 123.22, 126.4, 135.3, 136.5, 148.3. HREIMS: m/z: calcd
d¼18.0, 21.0, 57.1,
for C19H22N2 ([M]þ) 278.1783; found: 278.1786.
11. Khoshbin, M. S.; Ovchinnikov, M. V.; Mirkin, C. A.; Golen, J. A.; Rheingold, A. L.
Inorg. Chem. 2006, 45, 2603–2609.
12. Wilen, S. H.; Qi, J. Z.; Williard, P. G. J. Org. Chem. 1991, 56, 485–487; Tatibouet,
A.; Demeunynck, M.; Andraud, C.; Collet, A.; Lhomme, J. Chem. Commun. 1999,
161–162; Talas, E.; Margitfalvi, J.; Machytka, D.; Czugler, M. Tetrahedron:
Asymmetry 1998, 9, 4151–4156.
4.2.14. (ꢀ)-1,4,7,10-Tetramethyl-6H,12H-5,11-methanodibenzo-
[b,f][1,5]diazocine ((ꢀ)-3f)
Prepared according to the general procedure from 2,5-dime-
thylaniline 6f (606 mg, 5.00 mmol). The crude product was purified
by column chromatography (CH2Cl2/AcOEt 98:2) to afford pure
13. Satishkumar, S.; Periasamy, M. Tetrahedron: Asymmetry 2006, 17, 1116–1119.

6262
D. Didier et al. / Tetrahedron 64 (2008) 6252–6262
14. Lenev, D. A.; Lyssenko, K. A.; Golovanov, D. G.; Buss, V.; Kostyanovsky, R. G.
Chem.dEur. J. 2006, 12, 6412–6418.
15. Artacho, J.; Nilsson, P.; Bergquist, K.-E.; Wendt, O. F.; Wa¨rnmark, K. Chem.dEur.
J. 2006, 12, 2692–2701.
16. Lenev, D. A.; Lyssenko, K. A.; Golovanov, D. G.; Malyshev, O. R.; Levkin, P. A.;
Kostyanovsky, R. G. Tetrahedron Lett. 2006, 47, 319–321.
17. Kostyanovsky, R. G.; Kostyanovsky, V. R.; Kadorkina, G. K.; Lyssenko, K. A.
Mendeleev Commun. 2003, 111–113.
36. ESIMS of the reaction mixture after work up displayed the corresponding
[MþH]þ pseudomolecular ion, and 1H NMR spectrum revealed typical features
of the methanodiazocine bicycle, that is, an AB system (2JHendo,Hexo¼16.5 Hz) at
4.60 and 4.05 ppm, and a singlet of the NCH2N bridge at 4.25 ppm. In the acidic
medium, protonation converts Me2N group into a very strong acceptor, which
would deactivate the aromatic ring toward electrophilic substitutions. How-
ever, there is always some free amine in equilibrium with an ammonium
cation.
´
18. Prelog, V.; Wieland, P. Helv. Chim. Acta 1944, 27, 1127–1134.
19. Worlitschek, J.; Bosco, M.; Huber, M.; Gramlich, V.; Mazzotti, M. Helv. Chim. Acta
2004, 87, 279–291; Mihlbachler, K.; Kaczmarski, K.; Seidel-Morgenstern, A.;
Guiochon, G. J. Chromatogr., A 2002, 955, 35–52; Amanullah, M.; Mazzotti, M.
J. Chromatogr., A 2006, 1107, 36–45; Juza, M.; Mazzotti, M.; Morbidelli, M.
Trends Biotechnol. 2000, 18, 108–118; Mihlbachler, K.; De Jesus, M. A.;
Kaczmarski, K.; Sepaniak, M. J.; Seidel-Morgenstern, A.; Guiochon, G.
J. Chromatogr., A 2006, 1113, 148–161; Mihlbachler, K.; Seidel-Morgenstern, A.;
Guiochon, G. AIChE J. 2004, 50, 611–624; Araujo, J. M. M.; Rodrigues, R. C. R.;
Eusebio, M. F. J.; Mota, J. P. B. J. Chromatogr., A 2008, 1189, 292–301.
20. Sergeyev, S.; Diederich, F. Chirality 2006, 18, 707–712.
21. Miyahara, Y.; Izumi, K.; Ibrahim, A. A.; Inazu, T. Tetrahedron Lett. 1999, 40, 1705–
1708; Kubo, Y.; Ohno, T.; Yamanaka, J.-i.; Tokita, S.; Iida, T.; Ishimaru, Y. J. Am.
Chem. Soc. 2001, 123, 12700–12701; Allen, P. R.; Reek, J. N. H.; Try, A. C.; Crossley,
M. J. Tetrahedron: Asymmetry 1997, 8, 1161–1164; Kiehne, U.; Bruhn, T.; Schna-
kenburg, G.; Fro¨hlich, R.; Bringmann, G.; Lu¨tzen, A. Chem.dEur. J. 2008, 14,
4246–4255.
22. Wu, H.; Zhang, P.; Shen, Y.; Zhang, F.-R.; Wan, Y.; Shi, D.-Q. Synlett 2007, 336–338.
23. Wilcox, C. S. Tetrahedron Lett. 1985, 26, 5749–5752.
24. Sucholeiki, I.; Lynch, V.; Phan, L.; Wilcox, C. S. J. Org. Chem. 1988, 53, 98–104.
25. Li, Z.; Xu, X.; Peng, Y.; Jiang, Z.; Ding, C.; Qian, X. Synthesis 2005, 1228–
1230.
26. Ha¨ring, M. Helv. Chim. Acta 1963, 46, 2970–2982.
27. Jensen, J.; Wa¨rnmark, K. Synthesis 2001, 1873–1877.
28. Hansson, A.; Jensen, J.; Wendt, O. F.; Wa¨rnmark, K. Eur. J. Org. Chem. 2003,
3179–3188.
37. Havlik, M.; Kral, V.; Dolensky, B. Org. Lett. 2006, 8, 4867–4870; Valık, M.; Cejka,
J.; Havlik, M.; Kra´l, V.; Dolensky, B. Chem. Commun. 2007, 3835–3837; Hansson,
A.; Wixe, T.; Bergquist, K.-E.; Wa¨rnmark, K. Org. Lett. 2005, 7, 2019–2022; Pardo,
C.; Sesmilo, E.; Gutierrez-Puebla, E.; Monge, A.; Elguero, J.; Fruchier, A. J. Org.
Chem. 2001, 66, 1607–1611.
38. Solano, C.; Svensson, D.; Olomi, Z.; Jensen, J.; Wendt, O. F.; Wa¨rnmark, K. Eur. J.
Org. Chem. 2005, 3510–3517; Jensen, J.; Strozyk, M.; Wa¨rnmark, K. Synthesis
2002, 2761–2765; Bew, S. P.; Legentil, L.; Scholier, V.; Sharma, S. V. Chem.
Commun. 2007, 389–391; Kiehne, U.; Lu¨ tzen, A. Synthesis 2004, 1687–1695.
39. Didier, D.; Sergeyev, S. Tetrahedron 2007, 63, 3864–3869.
40. Wagner, E. C.; Eisner, A. J. Am. Chem. Soc. 1937, 59, 879–883.
41. Wagner, E. C. J. Org. Chem. 1954, 19, 1862–1881.
42. The synthesis of (ꢀ)-2e was very recently reported by Lu¨ tzen and co-workers,
Ref. 9. We thank Prof. A. Lu¨tzen and Dr. U. Kiehne for the helpful discussion
concerning the variation of conditions in this reaction.
43. It should be noted that in most cases the addition of reactants should be
performed slowly and upon cooling. An attempt to initiate the reaction be-
tween 4-iodoaniline and paraformaldehyde by addition of reactants to
CF3COOH in one portion without cooling resulted in an exothermic reaction
that produced a tarry, dark material and only traces of the desired Tro¨ger’s base
analogue.
44. Di- or tetramethyl derivatives (ꢀ)-3c–f should be more reactive toward elec-
trophilic substitution than (ꢀ)-1b. In a preliminary experiment, we confirmed
a much higher reactivity of (ꢀ)-3c compared to (ꢀ)-1b in the electrophilic
bromination. This study is currently in progress and will be published in due
course.
29. Jensen, J.; Strozyk, M.; Wa¨rnmark, K. J. Heterocycl. Chem. 2003, 40, 373–375.
30. Faroughi, M.; Scudder, M.; Turner, P.; Jensen, P.; Try, A. C. Acta Crystallogr., Sect.
E: Struct. Rep. Online 2007, 63, o1048–o1050; Faroughi, M.; Turner, P.; Try, A. C.
Acta Crystallogr., Sect. E: Struct. Rep. Online 2007, 63, o1045–o1047.
31. Salez, H.; Wardani, A.; Demeunynck, M.; Tatibouet, A.; Lhomme, J. Tetrahedron
Lett. 1995, 36, 1271–1274; Baldeyrou, B.; Tardy, C.; Bailly, C.; Colson, P.; Houssier,
C.; Charmantray, F.; Demeunynck, M. Eur. J. Med. Chem. 2002, 37, 315–322;
Tatibouet, A.; Demeunynck, M.; Lhomme, J. Synth. Commun. 1996, 26, 4375–
4395; Carree, F.; Pardo, C.; Galy, J.-P.; Boyer, G.; Robin, M.; Elguero, J. ARKIVOC
2003, 1–8.
45. Pirkle, W. H.; Welch, C. J. J. Liq. Chromatogr. 1992, 15, 1947–1955; Pirkle, W. H.;
Welch, C. J.; Lamm, B. J. Org. Chem. 1992, 57, 3854–3860.
46. Del Rio, A.; Piras, P.; Roussel, C. Chirality 2006, 18, 498–508.
47. Koscho, M. E.; Spence, P. L.; Pirkle, W. H. Tetrahedron: Asymmetry 2005, 16,
3147–3153; Koscho, M. E.; Pirkle, W. H. Tetrahedron: Asymmetry 2005, 16, 3345–
3351; Pirkle, W. H.; Welch, C. J. J. Chromatogr., A 1994, 683, 347–353; Del Rio, A.;
Hayes, J. M.; Stein, M.; Piras, P.; Roussel, C. Chirality 2004, 16, S1–S11.
48. Chiral Application Guide VI; Regis Technologies: Morton Grove, USA, 2006; p 15.
Available online at www.registech.com/chiral/.
49. Pirkle, W. H.; Gan, K. Z. J. Chromatogr., A 1997, 790, 65–71.
50. Pirkle, W. H.; Gan, K. Z.; Brice, L. J. Tetrahedron: Asymmetry 1996, 7,
2813–2816.
32. Didier, D.; Sergeyev, S. Eur. J. Org. Chem. 2007, 3905–3910.
33. Miller, T. R.; Wagner, E. C. J. Am. Chem. Soc. 1941, 63, 832–836.
34. Faroughi, M.; Try, A. C.; Turner, P. Acta Crystallogr., Sect. E: Struct. Rep. Online
2006, 62, o3893–o3894.
51. Hof, F.; Scha¨r, M.; Scofield, D. M.; Fischer, F.; Diederich, F.; Sergeyev, S. Helv.
Chim. Acta 2005, 88, 2333–2344.
35. Vande Velde, C. M. L.; Didier, D.; Blockhuys, F.; Sergeyev, S. Acta Crystallogr.,
Sect. E: Struct. Rep. Online 2008, 64, o538.
52. Doyle, M. P.; Van Lente, M. A.; Mowat, R.; Fobare, W. F. J. Org. Chem. 1980, 45,
2570–2575.